Hi, I have been exploring a PDB model that contains an RNA molecule with a ligand stuck on the 3' phosphate. To perform real-space refinement in Coot I needed to create a standard geometry CIF for the ligand and one for the linkage between the ligand and an Adenine. The CIF for the ligand in CCP4 and the PDBe has (in my opinion) serious defects, but with help from the masters at Global Phasing I was able to get a fine set of restraints from the Grade server.
To generate the linkage CIF I followed the very clear instructions at https://www2.mrc-lmb.cam.ac.uk/groups/murshudov/content/acedrg/acedrg.html to use acedrg. Long story short, one creates a one line description of the link and feeds that into acedrg along with the two relevant CIFs and Bob's your uncle. acedrg didn't work for me and I think I've uncovered two bugs that I had to work-around to get my file. I don't know who to report these bugs to -- hence this letter. I typed the command to cause acedrg to produce my link, but it failed saying (among other things) The component A.cif contains a deloc or aromatic bond. It needs to be kekulized. You could use acedrg to do kekulication of the bonds. Try getting A.cif from PDB/CCD and then use it as an input file to run acedrg to generate a cif file for the ligand description. e.g. You get the file A.cif from PDB/CCD and then, acedrg -c A.cif -o A_fromAcedrg ... The suggested second run of acedrg fails with the message (again truncated) The system contains atoms of the following elements P O C N H "H5' The input ligands/molecules contains metal or other heavier atoms Acedrg currently deals with ligands/molecules with following elements only C, N, O, S, P, B, F, Cl, Br, I, H There is no atom in A.cif from the CCP4 library with element "H5'. It does contain an atom with the odd name "H5' ". Knowing that many programs have problems with embedded spaces in names, and seeing no need for this embedded space, I changed the atom name and tried again. acedrg now completes the operation and writes out a new cif. This is bug #1. One might also consider the space in the name of this hydrogen atom to be a bug, but A.cif is a widely used file and it works for it principal purposes. Unfortunately, using the modified cif in the command for creating the link still fails. It appears that the kekulizer failed to kekulate. To investigate this kekulonic fault I examined the kekuleated cif and found that the "delocalized" bonds of the phosphate were not affected by the kekuliator. I attempted to manually kekulinate it by changing these bonds to a mixture of single and double bonds, with some trial and error required to get an acceptable mix, and created a cif that was now acceptably kekulinzed. This file was used by acedrg and I got the link I wanted. Bug #2 is the failure of acedrg's kekulinifier to remove all of the delocalized bonds. I hope this note makes it to acedrg's developers and they find it useful. It may also be helpful to others attempting this task (at least until the bugs are fixed). Dale Tronrud ######################################################################## To unsubscribe from the CCP4BB list, click the following link: https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=CCP4BB&A=1
